Projected stochastic primal-dual method for constrained online learning with kernels

Keyphrases

• Online Learning 100%
• Primal-dual Algorithm 100%
• Learning Kernels 100%
• Reproducing Kernel Hilbert Space 100%
• Decision Variables 50%
• Low-dimensional Subspace 50%
• Matching Pursuit 50%
• Vector-valued 50%
• Mean Convergence 50%
• Stochastic Gradient Descent 50%
• Constraint Violation 50%
• Saddle Point Problem 50%
• Nonlinear Constraints 50%
• Function Space 50%
• Constant Step Size 50%
• Supervised Learning 50%
• Constrained Problems 50%
• Stochastic Optimization 50%
• Primal-dual 50%
• Risk-aware 50%
• Dual Augmented Lagrangian 50%
• Representer Theorem 50%

Mathematics

• Stochastics 100%
• Hilbert Space 66%
• Step Size 66%
• Decision Variable 33%
• Optimality 33%
• Saddle Point 33%
• Mean Convergence 33%
• Lagrangian 33%
• Function Space 33%
• Solution Exists 33%
• Dimensional Subspace 33%